How Do We Know When an AV Is Safe Enough?

Introduction: The Safety Paradox

When I first started covering autonomous vehicles, I was struck by how much the public and industry alike were grappling with one central question: how do we know when an AV is safe enough to hit the road? This isn't just a technical issue; it's a deeply human one that touches on trust, responsibility, and the very nature of progress.

The Current Landscape

Today, there are no standardized tests for automated driving systems. While we require drivers to pass a test before they're allowed behind the wheel, there isn't an equivalent standard for autonomous technology. This gap has left companies like Waymo, Zoox, and Tesla to determine what constitutes safe performance on their own.

"Our top priority for this consortium is ADS competency, which will inform and accelerate a corresponding federal motor vehicle safety standard," said NHTSA Administrator Jonathan Morrison at the AI Horizons Summit in Pittsburgh.

This statement reflects a shift in thinking — that we can't rely on the industry alone to police itself. The National Highway Traffic Safety Administration (NHTSA) is beginning to recognize this, launching a new initiative called ASCEND to work with industry leaders and develop data-driven performance standards for autonomous driving systems.

The Challenge of Standards

Developing standards for AVs presents unique challenges. Unlike traditional vehicles where the Federal Motor Vehicle Safety Standards (FMVSS) provide a clear framework, self-driving cars lack many of the conventional features that these rules were built around — things like steering wheels, pedals, and mirrors.

This has created what I call an innovation paradox: we want to encourage rapid advancement in autonomous technology while simultaneously ensuring public safety. But without consistent standards, companies are left to create their own, leading to a patchwork of varying levels of performance and safety expectations.

Regulatory Progress and the ASCEND Initiative

What's encouraging is that regulators are finally moving toward a more structured approach. The ASCEND initiative between SAE and NHTSA aims to establish a baseline for how AVs should perform in real-world conditions. It's not just about testing the technology; it's about setting up a framework that will evolve with the technology.

"Ultimately, our ADS performance rulemaking will provide federal leadership in this space," Morrison explained. "By delivering an AV competency standard, we hope to address the burgeoning patchwork of inconsistent state and local level requirements..."">

This approach acknowledges that AV development isn't a linear process. As technology improves, so must the standards that govern it. It's a forward-looking strategy that recognizes the need for regulatory agility.

Industry Response and Innovation

The automotive industry is watching closely. While some companies are pushing for more relaxed regulations to accelerate deployment, others see value in stricter oversight. This tension is natural — after all, the goal of AV development is to create vehicles that can drive themselves better than humans, but only if they're safe enough to do so.

May Mobility's recent announcement to go public through a SPAC deal highlights this dynamic. The company, which operates autonomous Toyota Siennas in three cities, aims to be the first publicly traded robotaxi company in the U.S. Their "asset-light" approach emphasizes partnerships with existing vehicle manufacturers rather than building everything from scratch.

Real-World Testing and Performance

We've seen real-world testing efforts like Joby Aviation's 3,100-mile flight without a human pilot — not just a demonstration, but a practical application of autonomous technology. But we also have to look at failures, like Waymo's vehicle swept away during flash flooding in San Antonio, which reminded us that even advanced systems can encounter unforeseen challenges.

The key is not to stop innovation, but to ensure that testing and deployment happen responsibly. Every real-world test case adds valuable data — whether positive or negative — that helps refine safety protocols and standards.

Looking Ahead: The Road to Full Autonomy

As we move toward full autonomy, the question of safety becomes even more critical. We're not just talking about safer driving; we're talking about ensuring that machines make ethical decisions when faced with complex scenarios. This requires a level of standardization and transparency that hasn't been fully realized yet.

The path forward lies in collaboration between regulators, industry players, and the public. It's a delicate balance between encouraging innovation and protecting lives. We can't wait for perfect standards — we need standards that evolve with the technology, based on data from real-world deployments.

Until then, we'll continue to see a mix of cautious optimism and growing scrutiny as the industry works to build public trust in autonomous vehicles.